Anyone who has a handle on the U-frame stuffing issue:
I've read John K's rule-of-thumb guidelines on designing U-frames, and the pages of theory, etc and find it all practical.
However, the ongoing caveat about performance being contingent upon the u-frame being "properly stuffed" doesn't offer much direction for how to achieve that.
My first u-frame effort a while back was partially successful. Good bang for the buck . . . But something was not quite right, so I moved on to dipole alignments. I did not know at that time that the stuffing played a critical role; I simply took a stuff-to-taste approach, trying to make the peak go away. I probably was not getting the full benefit of the alignment. I'd like to try a U again, to get that efficiency.
Most DIY'ers aren't up to measuring group delay and such, but I CAN measure frequency response, Fs, and Qts of a completed system.
I know that data trends as stuffing increases will probably reverse at some point, particularly as the frame approaches being more of an aperiodic box than a transmission line. If I simply try to minimize the rear output as the graphs lead one to think, I'll likely over-stuff.
My question is this: As I try stuffing variations, is there any rule(s) of thumb, based on trends in Fs, Qts, or Fpeak (things I can readily measure) that will let me know I'm close? My ear will be the final judge, but it's much more fun if I can be in the right neighborhood first.
I hate to rule out U-frames until I've had the chance to evaluate one that is representative of "properly stuffed".
-- Mark